gpt-oss-20b on RTX 4090 (24GB) + AMD EPYC 7642 48-Core Processor (48c) + 252GB
MeasuredOn 2026-10-14, gpt-oss-20b Q4_K_M ran at 242.6 tok/s (chat-short) on an RTX 4090 (24GB) with llama.cpp. Run r_t-a9v_ipdit.
RTX 4090 (24GB) + AMD EPYC 7642 48-Core Processor (48c) + 252GB
suite suite-v1
cli 0.0.7
signedSgW66kVwx6…
Embed badgesubmitted Oct 14, 2026Workload results
| Workload | Backend | Model | decode tok/s | prefill tok/s | TTFT | p50 | p95 |
|---|---|---|---|---|---|---|---|
| chat-short | llama.cpp@0.6.0-dev (build 1, commit df18bee) | gpt-oss-20bQ4_K_M | 242.6tok/s | 593.8tok/s | 280ms | 4.1ms | 4.2ms |
| chat-long | llama.cpp@0.6.0-dev (build 1, commit df18bee) | gpt-oss-20bQ4_K_M | 231.2tok/s | 7,864.1tok/s | 405ms | 4.3ms | 4.3ms |
| concurrent-decode | llama.cpp@0.6.0-dev (build 1, commit df18bee) | gpt-oss-20bQ4_K_M | 239.6tok/s | — | — | 4.2ms | 4.2ms |
| agent-trace | llama.cpp@0.6.0-dev (build 1, commit df18bee) | gpt-oss-20bQ4_K_M | 236.1tok/s | 13,439.2tok/s | 150ms | 4.2ms | 4.3ms |
Reproduce on your machine
Start with this model and workload selection. This command does not capture the original artifact, backend version or server settings:
Match quantization, context capacity, caching and runtime settings for a useful comparison. Duration varies by workload and hardware; the published client may differ from this run's version. How it's measured.
Share this run
Copy the short link, a Reddit-ready summary, or a README badge. Every number is the measured result on this page, and each badge links back to it.
https://llm-speed.com/r/r_t-a9v_ipditRelated benchmarks
- More gpt-oss-20b benchmarks — every backend and rig that has run this model.
- More RTX 4090 (24GB) LLM benchmarks — every model measured on this hardware.
Provenance
- Run ID
- r_t-a9v_ipdit
- Fingerprint hash
- Public key
- SgW66kVwx64clisjWItmPvKz9Ijgd5DfuC5LS8Zmvtc=
- Received
- 2026-10-14 09:06:58